||The number of IP phone users is increasing. Thus, the number of errors in IP phone increases, and various factors cause the errors in a call control part. Identifying and analyzing the errors leads to improving in- terconnectivity between devices and providing good quality services. However, it requires a lot of effort to identify these errors because the identifying is done by visual observation. Therefore, we propose an automatic error identi- fication. In order to identify errors automatically, it needs training data. However, collecting training data in each environment is not easy because the collecting requires expert knowledge and an abundance of time. Therefore, we consider reusing training data. However, there is a possibility that identification accuracy decreases due to differences of data between environments. In this paper, we show the differences in SIP (Session Initiation Protocol) environment, and propose an automatic error identification that copes with the differences. We use Levenshtein distance in the identification method.